Accurate Identification of complex Land use and Land Cover Features using IRS (LISS III) Multispectral Image

Abstract

Land Use and Land Cover (LULC) is an assortment of activities executed by humans on to the land. The present study was carried out to evaluate supervised classification mechanisms for classification complex Land use and Land cover features using India Remote Sensing System-IRS (Linear Imaging Self-Scanning Sensor 3- LISS III) multispectral data. It showed that Artificial neural networks (ANN) fared better across all the land use and land cover classes with an overall accuracy of 88%. It also revealed that Maximum Likelihood (ML) and Support Vector Machine (SVM) classifier is prone to miss classification of pixels in one or more classes. Outcomes of the present study are comforting the competence of IRS (LISS III) multispectral data for the accurate mapping of complex land use and land cover features. Additionally, the ability of an ANN classifier in the classification of complex features using multispectral data was re-established in the present study.